Science o he To al En i onmen 922 (2024) 171210
A ailable online 26 Feb ua y 2024
0048-9697/© 2024 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-
nc-nd/4.0/).
Clima e change as a eiled d i e o mig a ion in Bangladesh and Ghana
Sa a Fe n´
andez
a
,
*
, Guadalupe A ce
b
, ´
Angela Ga cía-Alaminos
c
, Ignacio Cazca o
d
,
e
,
,
I˜
naki A o
a
Depa men o Applied & S uc u al Economics & His o y, Facul y o Economics and Business, Complu ense Uni e si y o Mad id, Campus de Somosaguas, 28223,
Pozuelo de Ala c´
on, Mad id, Spain
b
Escuela T´
ecnica Supe io de Ingenie ía Ag on´
omica y de Mon es y Bio ecnología, Uni e sidad de Cas illa-La Mancha (UCLM), Campus Uni e si a io, s/n, 02071
Albace e, Spain
c
Depa men o Economic Analysis and Finances, Uni e si y o Cas illa-La Mancha, Albace e, Spain
d
ARAID (A agonese Founda ion o Resea ch & De elopmen ), Za agoza, Spain
e
Ins i u o Ag oalimen a io de A ag´
on-IA2 (Uni e sidad de Za agoza-CITA), Depa amen o de An´
alisis Econ´
omico, Za agoza, Spain
Basque Cen e o Clima e Change, Leioa, Bizkaia, Spain
HIGHLIGHTS GRAPHICAL ABSTRACT
•Clima e d i e s o mig a ion in he
del as o Bangladesh and Ghana a e
analysed.
•The s udy is ca ied ou a he mic o
le el using he DECCMA da abase.
•Households do no iden i y en i on-
men al p essu es as he main cause o
mig a ion.
•Clima e shocks a ec ing economic se-
cu i y a e key d i e s o mig a ion in
del as.
•En i onmen al s ess emphasises he
occupa ion a iable as a d i e o
mig a ion.
ARTICLE INFO
Edi o : Jay Gan
JEL codes:
C25
O15
Q51
Q54
Q56
Keywo ds:
Fo ced mig a ion
Clima e change
Clima ic mig a ions
ABSTRACT
People li ing in del aic a eas in de eloping coun ies a e especially p one o su e he e ec s om na u al di-
sas e s due o hei geog aphical and economic s uc u e. Clima e change is con ibu ing o an inc ease in he
equency and in ensi y o ex eme e en s a ec ing he en i onmen al condi ions o del as, h ea ening he so-
cioeconomic de elopmen o people and, e en ually, igge ing mig a ion as an adap a ion s a egy. Clima e
change will likely con ibu e o wo sening en i onmen al s ess in del as, and unde s anding he ela ions be-
ween clima e change, en i onmen al impac s, socioeconomic condi ions, and mig a ion is eme ging as a key
elemen o planning clima e adap a ion. In his s udy, we use da a om mig a ion su eys and econome ic
echniques o analyse he ex en o which en i onmen al impac s a ec indi idual mig a ion decision-making in
wo del a egions in Bangladesh and Ghana. The esul s show ha , in bo h del as, clima ic shocks ha nega i ely
a ec economic secu i y a e signi ican d i e s o mig a ion, al hough he su eyed households do no iden i y
en i onmen al p essu es as he oo cause o he displacemen . Fu he mo e, en i onmen al impac s a ec ing
* Co esponding au ho .
E-mail add esses: [email p o ec ed] (S. Fe n´
andez), [email p o ec ed] (G. A ce), [email p o ec ed] (´
A. Ga cía-Alaminos), [email p o ec ed]
(I. Cazca o), [email p o ec ed] (I. A o).
Con en s lis s a ailable a ScienceDi ec
Science o he To al En i onmen
jou nal homepage: www.else ie .com/loca e/sci o en
h ps://doi.o g/10.1016/j.sci o en .2024.171210
Recei ed 3 No embe 2023; Recei ed in e ised o m 29 Janua y 2024; Accep ed 21 Feb ua y 2024
Science o he To al En i onmen 922 (2024) 171210
2
En i onmen al s ess
Adap a ion
Del a egions
ood secu i y and c op and li es ock p oduc ion a e also signi ican as e en s inducing people o mig a e, bu
only in Ghana. We also ind ha su e ing om en i onmen al s ess can in ensi y o educe he e ec s o so-
cioeconomic d i e s. In his sense, ad e se clima ic shocks may no only ha e a di ec impac on mig a ion bu
may also condi ion mig a ion decisions indi ec ly h ough he occupa ion, he educa ion, o he ma i al s a us o
he pe son. We conclude ha al hough clima e change and ela ed en i onmen al p essu es a e no pe cei ed as
key d i e s o mig a ion, hey a ec mig a ion decisions h ough indi ec channels (e.g., educing economic
secu i y o ein o cing he e ec o socioeconomic d i e s).
1. In oduc ion
Clima e change is a game-changing phenomenon in all sphe es o
human li e. La ge numbe s o people mig a e in olun a ily because o
clima e p essu es ha ei he a ec hei quali y o li e, hei sou ce o
income o bo h. The de ini ion o en i onmen ally induced mig a ion
p oposed by he In e na ional O ganisa ion o Mig a ion (IOM) (2007)
s a es ha “en i onmen al mig an s a e pe sons o g oups o pe sons
who, o compelling easons o sudden o p og essi e changes in he
en i onmen ha ad e sely a ec hei li es o li ing condi ions, a e
obliged o lea e hei habi ual homes, o choose o do so, ei he
empo a ily o pe manen ly, and who mo e ei he wi hin hei e i o y
o ab oad”. Acco ding o Bilak e al. (2016) an annual a e age o 21.5
million people had been o cibly displaced by wea he - ela ed sudden-
onse haza ds each yea since 2008, and he UNHCR (2022) highligh ed
ha nea ly 32 million displacemen s caused by wea he - ela ed haza ds
in 2022 ep esen a 41% inc ease compa ed o 2008 le els (es ima ed in
close o 23 million ha yea ), o which 98% we e caused by wea he -
ela ed haza ds such as loods, s o ms, wild i es and d ough s, acco d-
ing o he IDMC (2023).
Me a-analyses and e iews o he ela ionship be ween clima e
change, en i onmen al change and mig a ion can be ound in Ho mann
e al. (2020), Kaczan and O gill-Meye (2020), Beine and Jeuse e
(2021), Pigue e al. (2011), Wee asinghe (2021) and ob iously he IPCC
(2023). Wi hin hose s udies, and some o he s ha we e e speci ically
nex , e idence is p esen ed on how clima ic e en s lead o signi ican
changes (wa e sho age and d ough s, land deg ada ion a ec ing ood
p oduc ion and secu i y, see e.g. He mans and McLeman (2021), on
housing, ene gy and heal h see e.g. Mazhin e al. (2020); Palinkas
(2020); S ole e al. (2021)) ha may lead o mig a ion. In Janua y 2024
a me a- eg ession analysis o en i onmen al mig a ion li e a u e has
appea ed (Zhou and Chi, 2024), mainly e lec ing ha ac oss all he
global li e a u e, en i onmen al s esso s did no appea as impo an
p edic o s o (ou /in/ne ) mig a ion, wi h mixed e idence ending o
epo a bi mo e ou mig a ion.
Slow onse impac s o clima e change may lead o a ound 2.8% o he
popula ion in Sub-Saha an A ica, Sou h Asia, and La in Ame ica (i.e.
>143 million people) mo ing wi hin hei coun y o o igin by 2050
(Rigaud e al., 2018); and, a a wo ldwide le el, Mye s (2002) o ecas
a ound 200 million en i onmen al e ugees in 2050. Despi e wha hese
da a show, un il now, ew wo ks in he li e a u e ha e add essed cli-
ma ic ac o s as d i e s o mig a ion.
Since he las cen u y, se e al classi ica ions ha e ied o explain he
di e en de e minan s o mig a ion. One o he i s is ha o Lee (1966),
which dis inguishes ou g oups o ac o s: hose linked o he a ea o
o igin, hose linked o he a ea o des ina ion, obs acles, and pe sonal
ac o s. Se e al yea s la e , Yo imi su (1985) ca ies ou a classi ica ion
o he majo mig a ion de e minan s consis ing o ou ca ego ies: (1)
demog aphic cha ac e is ics o mig an s, (2) socioeconomic cha ac e -
is ics o mig an s, (3) socioeconomic cha ac e is ics o places o o igin
and des ina ion, and (4) ac o s accompanied by mig a ion. A e wa ds,
o he li e a u e has dis inguished be ween h ee ypes o de e minan s
explaining mig a ions: oo causes, p oxima e condi ions, and in e -
ening ac o s (Schmeidl, 1997). Roo causes include ac o s such as
po e y o popula ion p essu es; p oxima e condi ions ocus on human
igh s iola ions as well as e hical, ci il, o mili a y con lic s; and
in e ening ac o s e e o mig a ion ne wo ks o obs acles o mig a-
ion. Howe e , i should be no ed ha his classi ica ion is based on a
s udy mainly on e ugees and no on a comple e analysis o mig a ion o
speci ically o en i onmen ally induced mig a ion, so he e may be
o he ac o s ha ha e no been conside ed.
In his sense, he li e a u e ela ed o mig a ion has ied o dis in-
guish be ween olun a y and o ced mig a ion. Volun a y mig a ions
would be hose ha occu ou o a desi e o maximize hei wel a e,
while o ced mig a ions a e hose ha occu in esponse o some kind o
shock, such as wa s (Kuhn , 2019). Howe e , mos mig an s would be
loca ed somewhe e in be ween he wo ypes, nei he being o ced mi-
g an s in hei en i e y no olun a y mig an s en i ely (E dal and
Oeppen, 2018). In his ega d, he e is a need o mo e esea ch ha
analyses he d i e s o mig a ions no only a a heo e ical le el,
showing he hie a chy o de e minan s, which has no ye been es ab-
lished (Kuhn , 2019), bu also combined wi h empi ically d i en
esea ch ha helps ine- une he ac o o d i e s’ analyses based on
e idence.
In addi ion o he o ced and olun a y mig a ion dis inc ion, i
should be made a di e en ia ion be ween in e nal and ex e nal mig a-
ion, as in e nal mo emen s a e pa icula ly impo an in de eloping
coun ies. Speci ically, in e nal mig a ion in de eloping coun ies can
lead o posi i e change in bo h sending and ecei ing a eas, ei he
educing po e y a es o os e ing economic de elopmen (Deshingka
and G imm, 2004). Howe e , he de elopmen bene i s o in e nal
mig a ion end o a ise mos ly when people mo e olun a ily, bu no
when mig a ion is o ced by ex e nal elemen s (The Wo ld Bank, 2009),
so clima ic mig a ions a e a p oblem ha mus be assessed. The IPCC
(2020) de ines a clima e isk as “ he po en ial o ad e se consequences
o human o ecological sys ems, ecognizing he di e si y o alues and
objec i es associa ed wi h such sys ems”. The concep o isk is essen ial
o unde s anding he inc easingly se e e, in e connec ed and o en
i e e sible impac s o clima e change on ecosys ems, biodi e si y, and
human sys ems; and how o bes educe ad e se consequences o cu -
en and u u e gene a ions (IPCC, 2022b). Hel be g and Bonch-
Osmolo skiy (2011) p oposed ha a household is ulne able o any
isk associa ed wi h clima e change i he isk gene a es a loss o wel-
a e
1
ha pushes he household below a ce ain h eshold le el.
Vulne abili y is a unc ion o he na u e o he isk, exposu e and
sensi i i y o i , and adap a ion capaci y. Some o he mos exposed
egions o clima e change isks a e he del as in de eloping egions o
Asia and A ica. Low lying ele a ion o as ac s o land makes del as
highly exposed o sea-le el ise, among o he clima e change impac s
such as s o m su ges o saliniza ion (B own e al., 2018; Jin e al., 2018;
Nicholls e al., 2019). In hose wo ks, e idence is shown ha del as in
India and Bangladesh ha e some o he highes popula ion densi ies
globally, mainly de o ed o ag icul u al and ishing occupa ions ha
s ongly depend on he monsoon ain all condi ions wi h low-income
and subsis ence li elihoods in many cases (Laza e al., 2015).
1
Wel a e and Well-being a e usually employed as synonyms. Howe e , we
e e o well-being as a mul idimensional e m ha e e s o a s a e o heal h,
happiness and/o p ospe i y; while we employ wel a e as a mo e speci ic
concep ha applies o quan i iable well-being, assuming he classical economic
assump ion ha a highe le el o u ili y cu e signi ies a be e condi ion o he
economic agen (Maximo, 2016).
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
3
The e o e, he cha ac e is ics o hese del a egions make hem espe-
cially ulne able o he socioeconomic consequences o clima e change
(A o e al., 2019; Das e al., 2021). In ac , hei socioeconomic
ulne abili y hinde s hei adap a ion s a egies, which a e o en
insu icien o ace en i onmen al isks (FAO, 2022; Hossen e al., 2019;
Whi ehead e al., 2018). The e o e, he well-being o he communi ies o
such del as is endange ed by clima e change ac ing as a isk mul iplie
ha migh agg a a e o he p oblems in hese a eas (Ghosh e al., 2019;
Hossen e al., 2019), wi h he u al poo communi ies being he mos
a ec ed by clima e-change consequences ei he in India and Bangladesh
and sub-Saha an A ica (Ba ios e al., 2006; Pigue e al., 2011).
The social en i onmen in such del as is e y dynamic, making
mobili y a usual p ac ice. T adi ionally, economic mo i a ions we e he
main d i e o hese mig a ions. S ill, he ends o clima e change e -
ec s on hese egions poin o en i onmen al haza ds as one c ucial
d i e ha should be assessed (Jin e al., 2018; Sa a de Campos e al.,
2020; Samling e al., 2015). Gi en his si ua ion, he Del as, Vulne a-
bili y & Clima e Change: Mig a ion & Adap a ion (DECCMA) p ojec
was c ea ed o unde s and how clima e-change-d i en global and na-
ional mac o-economic p ocesses impac on mig a ion o men and
women in del as (DECCMA P ojec , 2022; Nicholls e al., 2019). The
p ojec iden i ies ou del as as especially ulne able a eas o clima e
change e ec s: he Bengal del a and he Mahanadi del a in India, he
Ganges-B ahmapu a-Meghna (GBM) del a in Bangladesh and he Vol a
del a in Ghana. We will ocus on he Vol a del a in Ghana and he
Bangladeshi side o he Ganges-B ahmapu a-Meghna (GBM) del a.
In ecen yea s, classi ica ions o mig a ion d i e s ha e begun o
include clima ic ac o s. Following he classi ica ion o he d i e s o
mig a ion by Van Hea e al. (2017), clima ic s ess as a push-d i e o
in olun a y mig a ion may ange om a p edisposing d i e in cases in
which mobili y is an adap a ion s a egy o a p ecipi a ing de e minan
when he displacemen is o ced in cases o li e- h ea ening haza ds ha
accele a es he decision o mig a ing (The Whi e House, 2021). En i-
onmen al and clima ic condi ions a e a ely a unique and di ec d i e
o mig a ion, bu hey can indi ec ly in luence mig a ion h ough hei
impac on o he social, economic, poli ical and demog aphic ac o s
unde lying hese mobili y decisions (Beine and Pa sons, 2015; Black
e al., 2011). Fo his eason, mig an s usually do no conside hei
decision as clima e-d i en, bu ins ead, hey pe cei e economic and
social ac o s as he main cause o hei mobili y (Adge e al., 2021;
Sa a de Campos e al., 2020). Howe e , clima ic shocks ha e been
p o en o be as impo an as educa ion, gende o ma i al s a us in
de e mining in e nal mig a ion in many coun ies (Abel e al., 2022).
Based on he abo e, his pape aims o ill his gap in he li e a u e (as
e.g. ound in Kuhn (2019)) by empi ically analysing he ex en o which
en i onmen al change isks play a ole in indi idual mig a ion decision-
making in ulne able del a egions. In his way, we in end o e eal i he
subjacen mo i a ion o he mig a ion is ela ed o clima e change
despi e he households do no explici ly iden i y i as he main eason o
he mig a ion (Adge e al., 2021; Sa a de Campos e al., 2020).
In addi ion, his esea ch uses a wide a ie y o en i onmen al
p essu e indica o s, which enables acing a he mic o le el he expo-
su e o di e en clima ic e en s ( loods, d ough s, e osion, salini y,
s o m su ges and cyclones) and hei e ec s on each household’s wel a e
and income. Scien i ic e idence claims ha he clima e change igge ed
by he ise in an h opogenic g eenhouse gases emissions is inc easing
he equency and in ensi y o hese kinds o ex eme wea he e en s
(IPCC, 2022a; NASA, 2021). The close an eceden o ou p oposal is he
wo k o Ho mann e al. (2019), which s udies he mo i a ion o u al-
u ban mig an s who mo ed om u al a eas in he Indian s a e o
U a akhand o i s capi al ci y. This s udy conside s he land and o es
co e changes a ound he chosen illages as a possible en i onmen al
d i e o he mig a ions, which is buil a he meso-le el using a
geog aphic in o ma ion sys em analysis o land co e changes. In ou
assessmen , we wo k wi h mic o-le el clima e indica o s wi h a high
le el o de ail, bo h in he a ie y o clima e e en s o which he
household is exposed, bo h in he e ec s o hese e en s on he house-
hold’s wel a e. Mo eo e , he da abase used in ou s udy p o ides in-
dica o s o en i onmen al s ess bo h in objec i e and subjec i e e ms.
In his way, ou model conside s he pe cep ion o he household con-
ce ning he impac s o clima e change phenomena on i s li es yle, which
migh be a ele an de e minan in he mig a ion decision.
The e o e, he main con ibu ions o his wo k a e h ee old. Fi s ,
using a ep esen a i e sample o del as o Bangladesh and Ghana, his
pape analyses he s ill unde explo ed clima e d i e s. Secondly, he
a emp o examine he e ec o hese di e en d i e s on wo ulne -
able del as wi h di e en cha ac e is ics, which allows us o ca y ou a
compa a i e analysis be ween he wo a eas. Finally, his pape ies o
cla i y whe he he mo i a ion o mig a ion is ela ed o clima ic ac-
o s, e en i households do no iden i y i as such. The ole o en i on-
men al s ess as mode a ing e ec making use o in e ac ion a iables
wi h mo e commonly s udied socioeconomic a iables esul s ele an
in he inal explana o y model.
2. Me hods and da a
Da a is e ie ed om quan i a i e su eys ca ied ou in 2016 as
pa o he p ojec Del as, Vulne abili y and Clima e Change: Mig a ion
and Adap a ion (Sa a de Campos and Adge , 2021). These su eys
add ess di e en issues such as he ci cums ances unde which he de-
cision o mig a e is aken, he condi ions unde which mig a ion is mo e
o less likely o be a success ul adap a ion o clima e change, o he
ac o s ha impede o acili a e success ul mig a ion, among o he s.
2
To collec he in o ma ion, he su eys we e ansla ed in o he main
language o he e i o ies and ca ied ou by local people. The egions in
which he s udy was ca ied ou a e ou del a egions selec ed as
ulne able o clima e change by he DECCMA p ojec (Sa a de Campos
and Adge , 2021): he la ges del a in he wo ld (Ganges-B ahmapu a-
Meghna (GBM) in Bangladesh), wo medium-sized del as (Indian Bengal
Del a -pa o he GBM - and Mahanadi in India), and a small-sized del a
(Vol a in Ghana). Fo easons o da a a ailabili y in en i onmen al s ess
and mo i a ion o he mig a ion ques ions, ou analysis has been ca ied
ou only o he del as o Ghana and Bangladesh. The e o e, his allows
us o s udy 2 del as ha ha e di e en cha ac e is ics and belong o
di e en geog aphical a eas: a e y la ge del a (in Bangladesh) and a
ela i ely small one in Ghana. Analysing ulne able del as wi h di e en
cha ac e is ics allows o conside scale, geog aphic se ings, and a ying
d i e s in ou analysis. Each del a s udy a ea has been delimi ed ac-
co ding o he i e-me e ele a ion con ou line o ocus a en ion on he
coas al p ocesses and haza ds linked o sea-le el ise (Laza e al., 2015).
Thus, ou sample size is inally 1328 households o Bangladesh and
Table 1
Rela ion o he mig an wi h he household head in hose households engaged in
mig a ion (pe cen ages o o al mig an households).
Bangladesh Ghana
Pa ne 124 (30.69%) 2 (0.47%)
Ma ied child 117 (28.96%) 44 (10.38%)
Unma ied child 106 (26.24%) 205 (48.35%)
Pa en 2 (0.49%) 152 (35.85%)
B o he /sis e 54 (13.37%) 3 (0.71%)
B o he -in-law/sis e -in-law 1 (0.25%) 18 (4.24%)
O he ela i es 0 0
Non- ela i es 0 0
Don’ know 0 0
To al 404 (100%) 424 (100%)
Sou ce: own elabo a ion wi h da a e ie ed om DECCMA 2016 da abase
(Sa a de Campos and Adge , 2021).
2
Fo mo e in o ma ion on he opics co e ed in he su ey, as well as
me hodological aspec s see: Sa a de Campos and Adge , 2021 and DECCMA
P ojec .
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
4
1300 households o Ghana.
3
The ques ionnai es a e answe ed by he pe son sel -iden i ied as
household head.
4
In households engaged in mig a ion (wi h a leas one
membe ha has mig a ed), a iables ela ed o indi iduals’ cha ac-
e is ics e e o he mig an ’s (i.e., age, occupa ion, ma i al s a us, e c.).
In households no engaged in mig a ion, he esponses e e o he
household head. This cons i u es a limi a ion o ou model ega ding
con ol a iables ela ed o he indi idual’s cha ac e is ics. Table 1
shows he ela ionship be ween he household head and he mig an in
hose households engaged in mig a ion. In Ghana, he household head
has a son/daugh e -pa en ela ionship wi h he mig an , while in
Bangladesh he pe sons s aying as he household head is mo e o en he
pa ne o he son/daugh e o he mig an , bu no he pa en .
En i onmen al s ess ques ions ackle a a ie y o clima ic e en s
such as lood, d ough , e osion, and saliniza ion, enqui ing bo h abou
hei magni ude and p obabili y. The ques iona y does no e ie e in-
o ma ion abou aspi a ions and desi es among he possible d i e s o
mig a ion. In consequence, ou de ini ion o mig a ion d i e s is aligned
wi h ha o Van Hea e al. (2017), who de ine hem as s uc u al ele-
men s ac ing as ex e nal o ces ha in luence mobili y. The mig a ion
pa e ns epo ed in he ques iona y a e bo h pe manen and empo a y,
which is ele an as bo h kinds o esponses appea as adap a ion s a-
egies in impac ed communi ies (Sa a de Campos e al., 2020), and
empo a y mig a ion should no be uled ou om his kind o analysis
(Abel e al., 2022; Boh a-Mish a e al., 2014; Joa de and Mille , 2013).
The empi ical s a egy consis s o wo pa s: a desc ip i e analysis o
he samples and an econome ic analysis. The econome ic model used is
a p obi model ha will allow us o analyse he p obabili y o mig a ion
and he ela i e in luence o each explana o y a iable on his decision.
Following G eene (2003), Eqs. (1) and (2) expose he analy ical o m o
he model:
y*
i=x
′
i,ENV β1+x
′
i,CTRLβ2+
ε
i,
ε
i∼N[0,1](1)
yi=1i y*
i>0,0o he wise (2)
whe e he la en a iable y*
i is de ined as he p opensi y o indi iduals o
mig a e, and i i exceeds a ce ain h eshold, he dependen a iable yi
will ake he alue 1 o 0 o he wise. The independen a iables a e
classi ied in o d i e s ela ed o en i onmen al s ess (x
′
i,ENV) and con ol
a iables (x
′
i,CTRL).
Speci ically, wo di e en p obi models will be used, and, he e o e,
wo dependen a iables will be conside ed. The i s one is mig a ion
o economic easons (mig aeco). This a iable akes he alue 1 i he
indi idual epo s mig a ing o seek employmen , housing p oblems,
deb p oblems o loss o income, and 0 o he wise. The second a iable is
mig a ion o social o amily easons (mig asoci ami), which will ake a
alue o 1 i he mig an epo s seeking educa ion, ma iage, amily
obliga ions, heal h ca e o social and/o poli ical p oblems as he eason
o mig a ing. On he o he hand, he ec o s x con ain he di e en
en i onmen al s ess and con ol a iables, he de ini ion o which can
be ound in Table 2. The gende a iable (GEN) is no in oduced in he
analysis o Bangladesh. The eason is ha male mig a ion domina es in
his coun y, wi h 94% o mig an s being men. The e o e, i seems clea
ha gende is signi ican in mig a ion in Bangladesh, bu ou aim is o
look a u he ela ions ha could be dis o ed by his ea u e o he
sample. In he case o Ghana, 52% o he mig an s we e men, which
implies ha ing a mo e gende -balanced mig a ion ha allows us
conside ing gende as a sui able con ol a iable. Tables A.1 and A.2 in
Appendix A show desc ip i e s a is ics and co ela ion ma ix espec-
i ely. In addi ion, Va iance In la ion Fac o es s ha e been ca ied ou
o check he p oblems o mul icollinea i y, which sa is y he econo-
me ic equi emen s.
Fo a deepe analysis, in addi ion o s udying how en i onmen -
ela ed d i e s in luence he decision o mig a e, i is in e es ing o
s udy how en i onmen al ac o s can a ec he o he d i e s and hus
indi ec ly in luence he decision o mig a e (Black e al., 2011). To do
his, a p incipal componen analysis (PCA) was ca ied ou o he
Table 2
Desc ip ion o en i onmen al s ess and con ol a iables.
Meaning
En i onmen al s ess a iables
Housing (HOU) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on housing; and 0 o he wise.
Ecosecu i y (ECO) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on economic secu i y; and 0 o he wise.
C op (CRO) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on c op/li es ock disease; and 0 o he wise.
Wa e (WAT) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on d inking wa e ; and 0 o he wise.
Foodsecu i y (FSE) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on ood secu i y; and 0 o he wise.
Heal h (HEA) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed ha looding, d ough , e osion, salini y, s o m su ges, o cyclone had mode a e o high
nega i e impac s on he household’s heal h; and 0 o he wise.
Con ol a iables
Pe manen job (PJO) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed pe manen wo k; and 0 o he wise.
Age (AGE) Indi idual’s age
Educa ion (EDU) Bina y a iable (1-0) wi h a alue o 1 i he indi idual indica ed seconda y o highe educa ion, such as uni e si y; and 0 o he wise.
Ma i al (MAR) Bina y a iable (1-0) wi h a alue o 1 i he indi idual s a ed being cu en ly ma ied; and 0 o he wise.
Gende (GEN) Bina y a iable (1-0) wi h a alue o 1 i he indi idual is a male; and 0 o he wise.
Occupa ion (OCC) Ca ego ical a iable anging om 1 o 20 depending on he occupa ion indica ed by he indi idual.
1: C op a me , 2: Li es ock a me , 3: Fish/sh imp a me , 4: Fishing, 5: Regula sala ied employee, 6: Small business owne , 7: Cons uc ion wo ke , 8:
Fac o y wo ke , 9: Domes ic employee, 10: T ade , d essmake / ailo , 11: T anspo wo ke (i.e. ickshaw pulle , axi d i e ), 12: Hawke , 13: Gua d/
ga dene , 14: Money lende , 15: Unpaid home ca e , 16: Unemployed, 17: S uden , 18: Re i ed, 19: O he , 20: Don’ know.
Sou ce: own elabo a ion.
3
The su ey conside ed as household a g oup o people li ing in he same
dwelling and sha ing meals and/o expenses.
4
The su ey conside ed as household head he pe son who has he mos
au ho i y and esponsibili y o household a ai s.
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
5
independen a iables conside ed as en i onmen al d i e s. P incipal
componen analysis is a adi ional s a is ical me hod (Ho elling, 1933)
ha con e s a se o a iables ha migh be co ela ed in o linea ly
unco ela ed a iables ha a e called p incipal componen s. This
me hod ex ac s he dominan pa e ns and displays hem in e ms o
componen sco es and loadings (Wold e al., 1987). The sco es co e-
spond o he ans o med alues assigned o he a iables, while he
loadings will be used o ob ain he componen sco e by mul iplying hem
by each o iginal s anda dized a iable. To decide on he numbe o
p incipal componen s o keep, we will ollow Kaise ’s ule (Kaise ,
1960). Acco ding o his c i e ion, all componen s wi h an eigen alue
g ea e han uni y should be e ained. A e ha , he selec ed compo-
nen s o ac o s a e no malized by conside ing he minimum and
maximum alue o each ac o :
ac o −minimum
maximum −minimum(3)
Based on he p incipal componen s ob ained, an en i onmen al
s ess a iable was gene a ed. I was be in e ac ed wi h he classical
d i e s o mig a ion (con ol a iables) and hese in e ac ions we e
included as independen a iables in he p obi models. This ype o
a iables allows us o know i en i onmen al s ess in ensi ies o educes
he e ec o o he mig a ion de e minan s. The e o e, he in e ac ions o
see he mode a ing e ec o en i onmen al s ess on he con ol a ia-
bles— he classical mig a ion d i e s—a e in oduced in Eq. (1), leading
o he ollowing exp ession:
y*
i=x
′
i,ENV β1+x
′
i,CTRLβ2+x
′
i,ENVSTRESSx
′
i,CTRLβ3+
ε
i,
ε
i∼N[0,1](4)
whe e he e is an addi ional independen a iable (x
′
i,ENVSTRESS) ha e-
e s o he en i onmen al s ess a iable gene a ed wi h he PCA ech-
nique.
3. Resul s and discussion
3.1. Desc ip i e analysis
This sec ion is de o ed o he cha ac e iza ion o he sample in
Bangladesh and Ghana, di iding he obse a ions in o wo g oups o
each coun y: one o households in ol ed in mig a ion (wi h somebody
who has mig a ed) and ano he one o households no in ol ed in
mig a ion (no-one has mig a ed). In he i s case, he esponses gi en by
he pe son answe ing he su ey ( he sel -designa ed as household-head)
e e o he mig an cha ac e is ics, while in he second case he answe s
e e o he household head cha ac e is ics.
S a ing wi h he occupa ion
5
(OCC) by sec o (Fig. 1), in
Bangladesh, mig an s in he sample a e mainly occupied as egula
sala ied employees (37.1%), ollowed by cons uc ion wo ke s (13.8%),
ac o y wo ke s (7.5%) and small business owne s (7.3%), while he
p edominan occupa ions among non-mig an s a e small business
owne s (19.0%), c op a me s (15.2%), egula sala ied employees
(11.3%) and unpaid home ca e s (9.1%). Compa ing he pa icipa ion o
each occupa ion among bo h sub-samples, i makes sense ha occupa-
ions ela ed o non-mobile asse s -such as c op cul i a ion o owning a
small business- ha e lowe pa icipa ion among he mig an s. Acco ding
o Be nzen e al. (2019), ag icul u al and sel -employmen occupa ions
imply a highe eluc ance o mig a e because hese ac i i ies equi e
ce ain in es men s in he communi y, which in u al Bangladesh is
usually o ganized a ound managing asse s such as land and local busi-
ness (Ish iaque and Ullah, 2013). In addi ion, i is obse ed ha a e
mig a ion, indi iduals occupy jobs wi h be e condi ions ( egula
sala ied employees, cons uc ion, o ac o y wo ke s).
In he case o Ghana, he e a e mo e simili udes be ween mig an s
and non-mig an s han in Bangladesh. In bo h g oups, ade , d ess-
make / aylo is he occupa ion wi h highe pa icipa ion, accoun ing o
abou 25% o he o al in bo h cases, ollowed by egula sala ied em-
ployees and c op a me s, which sum up o >24% o bo h g oups, wi h
he di e ence o c op a me s being mo e equen in he non-mig an s
g oup in a simila way as in Bangladesh. In Ghana, unemployed people
cons i u e 8% o he mig an s in ou sample, while hei pa icipa ion
among he non-mig an s is jus 2.3%.
Table 3 displays an addi ional cha ac e iza ion o he di e ences
be ween he mig an and non-mig an households in he sample. I
shows he esul s o a - es de eloped o de e mine whe he he a e age
o he en i onmen al-haza d a iables in he model di e s o no be-
ween hese wo collec i es. Looking a hese esul s, he e a e clea
di e ences -e en unde 1% o signi icance le el in mos o he cases–
be ween he a e ages o each a iable o households in ol ed and no
in ol ed in mig a ion in Ghana, ega dless o he mo i a ion behind he
displacemen . In he case o Bangladesh, he wo g oups di e signi i-
can ly conce ning he model a iables in a e age e ms when looking a
socio- amilia mig a ions (excep o he d inking-wa e a iable –WAT-
). Howe e , when aking jus he sub-sample ela ed o economically-
mo i a ed mig a ions and when wo king wi h he whole sample
wi hou conside ing he eason o he displacemen , jus he housing
(HOU) and c op/li es ock disease (CRO) a iables show a di e en
a e age o households in ol ed and no in ol ed in mig a ion unde a
10% o he signi icance le el.
Focusing now on he ea u es o he mig an s in he sample, Figs. 2
and 3 show he cha ac e is ics o he des ina ion chosen and he eason
behind his elec ion. Acco ding o Fig. 2, u al des ina ions and nea by
se lemen s a e mo e equen in he sample o Ghana han in he one
o Bangladesh, while in e na ional mig a ion is no o iously highe
among he indi iduals su eyed in he Asian coun y. Majo ci ies and
dis ic capi als a e s onge a ac o s in ou sample o Ghana, bu
egional capi als ha e a simila pa icipa ion o a ound 12% in bo h
coun ies. Looking a Fig. 3, he pa e n behind he easons o choose a
speci ic des ina ion is e y simila among bo h samples, wi h a clea
p edominance o ha ing amily membe s in he a ea (>60% o he cases
in bo h coun ies), ollowed by ha ing iends he e (accoun ing o a
27% o he esponses in bo h samples). This ein o ces he idea o human
links behind de e minan d i e s o some o he decisions behind he
mig a ion p ocess (Ackah and Med ede , 2012; Boas, 2020; Ma in
e al., 2014).
3.2. Mig a ion d i e s: econome ic esul s
Table 4 shows he esul s o he p obi model shown in Eqs. (1) and
(2) aking economic and socio- amilia mig a ions in Bangladesh as
dependen a iables. Speci ically, he e a e h ee models o each ype o
mig a ion, whe e he a iables ha e been scaled o check he obus ness
o he esul s. The i s es ima ion o each model conside s only hose
en i onmen al s ess d i e s (x
′
i,ENV in Eq. (1)) ha a e mainly ela ed o
economic losses. The second es ima ions add he en i onmen al s ess
d i e s (x
′
i,ENV in Eq. (1)) ha a e mainly ela ed o heal h and basic
needs. Finally, he hi d model in oduces all he con ol a iables
(x
′
i,CTRL in Eq. (1)).
On he one hand, only clima ic shocks ha nega i ely a ec houses
(HOU) could be a d i e o economic mig a ion in Bangladesh, al hough
he e ec is only signi ican in he second model. This e idence is
in ui i e since i he housing has been physically damaged, he need o
look o an al e na i e in o he loca ions is p essing, a ou ing mig a ion
(Mye s e al., 2008). On he o he hand, clima ic shocks ha nega i ely
a ec he economic secu i y (ECO) o households in Bangladesh inc ease
socio- amily mig a ion.
In ela ion o he con ol a iables, we obse e ha no ha ing a
pe manen job (PJO) dis a ou s bo h ypes o mig a ion. Despi e he
highe ulne abili y o clima e e en s, low-income households migh be
5
Pos -mig a ion occupa ion.
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
6
less likely o mig a e due o hei lack o esou ces. The cos s in ol ed in
long-dis ance mig a ion a e o en high and ou o he budge o poo
households (Ka iki, 2011). Age is a key ac o in he decision o mig a e.
In line wi h he esul s ob ained in he li e a u e (Dus mann and Oka-
enko, 2014; an Dalen e al., 2005), a nega i e link is shown wi h bo h
ypes o mig a ion: younge indi iduals a e mo e willing o mig a e. In
gene al, i is ound ha mig an s a e he sons o he conside ed house-
hold head, ei he ma ied o unma ied, being also ele an he mig a-
ion o he couple (see Table 1).
In addi ion, he a iable ela ed o educa ion (EDU) is posi i e and
signi ican in he economic mig a ion model, showing ha he highe
he le el o educa ion, he highe he p obabili y o emig a ing o
economic easons. In his sense, people wi h highe educa ional a ain-
men ha e mo e ans e able asse s, which enables hem o mig a e wi h
highe chances o inding an income sou ce in o he place (Be nzen
e al., 2019). Rega ding he ma i al s a us (MAR), i is a signi ican
a iable in bo h kind o mig a ions, as ou esul s show ha being
ma ied dec eases he p obabili y o mig a ing. In he case o economic
mig a ion, women who mig a e independen ly a e o en unma ied,
di o ced o widowed, as ma ied women a e discou aged o mig a ing
Fig. 1. Occupa ion o mig an s (households in ol ed in mig a ion) and non-mig an s (households no -in ol ed in mig a ion) in Ghana and Bangladesh.
No e: he numbe s in b acke s co espond o he alue associa ed wi h each occupa ion in he a iable (see Table 2 o a iable desc ip ion).
Sou ce: own elabo a ion wi h da a e ie ed om DECCMA 2016 da abase (Sa a de Campos and Adge , 2021).
Table 3
-Tes o compa e mig an and non-mig an sub-samples in a e age e ms o he en i onmen al-haza d a iables in he model.
Bangladesh Ghana
A e age mig a ion A e age no mig a ion -Tes p-Value A e age mig a ion A e age no mig a ion -Tes p-Value
Economic mo i a ion
Housing (HOU) 0.619 0.568 −1.749 0.04** 0.38 0.291 −3.431 0.0003***
Ecosecu i y (ECO) 0.43 0.396 −1.151 0.1249 0.374 0.192 −7.459 0***
C op (CRO) 0.177 0.142 −1.619 0.05* 0.226 0.103 −6.101 0***
Wa e (WAT) 0.285 0.3 0.534 0.705 0.189 0.119 −3.547 0.0002***
Foodsecu i y (FSE) 0.26 0.266 0.212 0.584 0.289 0.163 −5.501 0***
Heal h (HEA) 0.226 0.216 −0.405 0.3428 0.129 0.092 −2.186 0.014**
Socio- amilia mo i a ion
Housing (HOU) 0.649 0.568 −2.352 0.009*** 0.471 0.291 −5.456 0***
Ecosecu i y (ECO) 0.496 0.396 −2.899 0.0019*** 0.411 0.192 −7.281 0***
C op (CRO) 0.217 0.142 −2.951 0.0016*** 0.257 0.103 −6.259 0***
Wa e (WAT) 0.324 0.3 −0.767 0.2215 0.243 0.119 −4.904 0***
Foodsecu i y (FSE) 0.324 0.266 −1.861 0.0315** 0.379 0.163 −7.486 0***
Heal h (HEA) 0.294 0.216 −2.633 0.0043*** 0.171 0.092 −3.559 0.0002***
Gene al
Housing (HOU) 0.609 0.568 −1.481 0.0694* 0.374 0.291 −3.267 0.0006***
Ecosecu i y (ECO) 0.423 0.396 −0.952 0.1705 0.365 0.192 −7.269 0***
C op (CRO) 0.169 0.142 −1.319 0.0937* 0.215 0.103 −5.715 0***
Wa e (WAT) 0.288 0.3 0.429 0.6659 0.182 0.119 −3.260 0.0006***
Foodsecu i y (FSE) 0.258 0.266 0.313 0.623 0.279 0.163 −5.216 0***
Heal h (HEA) 0.226 0.216 −0.403 0.3434 0.125 0.092 −1.962 0.025**
*** p<0.01, ** p<0.05, * p<0.1.
Sou ce: own elabo a ion wi h da a e ie ed om DECCMA 2016 da abase (Sa a de Campos and Adge , 2021).
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
7
on hei own by household esponsibili ies and mo he hood (E e sen
and an de Gees , 2020). I is also expec ed ha single men a e mo e
p one o ealloca e since hey a e also expec ed o be he b eadwinne s
and a e he e o e expec ed o main ain he same occupa ional p o ile
(Kanaiaupuni, 2000). Occupa ion (OCC) is he only con ol a iable ha
is no signi ican in he wo models. The eason migh be ha i s in lu-
ence on bo h kind o mig a ion is al eady co e ed by o he a iables
such as ha ing a pe manen job (PJO) o ha ing highe educa ion (EDU),
whose co ela ions wi h occupa ion a e −0.221 and −0.14, espec i ely
(see Table A.2 in Appendix A).
As explained be o e, he gende a iable (GEN) is no in oduced in
he eg ession. The explana o y analysis shows ha he as majo i y o
he mig an s in Bangladesh a e men (94% o he mig an in he sample),
which implies a s ong in luence o he gende on he p obabili y o
mig a e ha could mask o he e ec s i in oduced in he model.
The esul s o Ghana a e shown in Table 5. Fi s ly, clima ic e en s
a ec ing he household (HOU) seem o inc ease he likelihood o socio-
amily mig a ion in Ghana. Losing he home o ha ing i se e ely
damaged c ea es a p oblem a he household le el, whe e he amily
head will conside mig a ing o maximize he wel a e o he amily.
Rega ding ood secu i y (FSE), his appea s o be a key d i e o socio-
household mig a ion in Ghana, showing ha clima e e en s ha
comp omise ood secu i y inc ease he likelihood o socio-household
mig a ion. This esul is simila o ha ob ained by Rademache -
Schulz e al. (2014), also o Ghana, whe e hey highligh ed ha
mig a ion is used as a s a egy o deal wi h ad e se clima ic e en s ha
h ea en ood secu i y.
In addi ion, he economic secu i y a iable (ECO) and he a iable
ela ed o c op and li es ock s a us (CRO) a e posi i e and signi ican in
almos all he models. On he one hand, when a clima e e en s ongly o
mode a ely a ec s household economic secu i y (ECO), he p obabili y
o mig a ing inc eases, simila o wha was ound o Bangladesh in he
case o socio- amilia mig a ion. On he o he hand, when c ops and
li es ock a e nega i ely a ec ed by a clima e shock (CRO), bo h ypes o
mig a ion a e a ou ed. Clima ic e en s a ec ing ag icul u e a a sub-
dis ic le el a e likely o weaken isk-sha ing ne wo ks and hinde
oppo uni ies o employmen , inc easing he mo i a ion o mig a e
(G ay e al., 2020).
Rega ding con ol a iables, as i was he case o Bangladesh eco-
nomic mig a ion, educa ional le el (EDU) has a posi i e in luence in
bo h kind o mig a ions in Ghana. In a simila way as in Bangladesh, age
has an in e se ela ionship (i.e. people mig a ing a e ela i ely young,
ypically he sons/daugh e s o he household head, being in his case o
Ghana mainly he unma ied ones, and in e es ingly also in some cases
he conside ed a he o he household head). Looking a he gende
(GEN), being a woman inc eases he p obabili y o mig a ing o bo h
Fig. 2. Des ina ion o he mig an s in Ghana and Bangladesh.
Sou ce: own elabo a ion wi h da a e ie ed om DECCMA 2016 da abase (Sa a de Campos and Adge , 2021).
Fig. 3. Reasons o choose des ina ion o mig an s in Ghana and Bangladesh.
Sou ce: own elabo a ion wi h da a e ie ed om DECCMA 2016 da abase (Sa a de Campos and Adge , 2021).
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
8
Table 4
P obi model esul s o economic and socio- amily mig a ion in Bangladesh.
mig aeco mig asoci ami
(I) (II) (III) (I) (II) (III)
Housing (HOU) 0.037 0.054* 0.042 0.036 0.035 0.026
(0.027) (0.029) (0.035) (0.026) (0.028) (0.033)
Ecosecu i y (ECO) 0.004 0.014 0.04 0.040 0.041 0.078**
(0.029) (0.030) (0.036) (0.028) (0.028) (0.033)
C op (CRO) 0.044 0.060 0.038 0.062 0.055 0.018
(0.040) (0.042) (0.050) (0.038) (0.039) (0.041)
Wa e (WAT) −0.038 −0.017 −0.039 −0.001
(0.033) (0.042) (0.030) (0.040)
Foodsecu i y (FSE) −0.035 −0.019 −0.002 −0.012
(0.034) (0.046) (0.032) (0.041)
Heal h (HEA) 0.005 0.029 0.053 0.065
(0.037) (0.049) (0.037) (0.045)
Pe manen job (PJO) 0.163*** 0.136***
(0.033) (0.030)
Age (AGE) −0.015*** −0.012***
(0.001) (0.001)
Educa ion (EDU) 0.100*** 0.013
(0.032) (0.029)
Ma i al (MAR) −0.262*** −0.249***
(0.047) (0.049)
Occupa ion (OCC) 0.005 0.003
(0.003) (0.003)
Obse a ions 1328 1328 1086 1183 1183 949
Pseudo R
2
0.003 0.005 0.2153 0.011 0.013 0.2128
X
2
4.77 8.16 239.12 13.50 16.30 180.16
Ma ginal e ec s. Robus s anda d e o s in pa en heses.
Sou ce: own elabo a ion.
***
p <0.01.
**
p <0.05.
*
p <0.1.
Table 5
P obi model esul s o economic and socio- amily mig a ion in Ghana.
mig aeco mig asoci ami
(I) (II) (III) (I) (II) (III)
Housing (HOU) 0.033 0.027 0.036 0.102*** 0.082** 0.086**
(0.031) (0.033) (0.036) (0.034) (0.035) (0.037)
Ecosecu i y (ECO) 0.168*** 0.152*** 0.191*** 0.136*** 0.079* 0.066
(0.037) (0.041) (0.046) (0.043) (0.047) (0.050)
C op (CRO) 0.116** 0.106** 0.111** 0.131*** 0.096* 0.117**
(0.045) (0.046) (0.050) (0.051) (0.053) (0.056)
Wa e (WAT) 0.029 0.029 0.034 0.027
(0.045) (0.048) (0.047) (0.047)
Foodsecu i y (FSE) 0.030 0.035 0.111** 0.140***
(0.044) (0.049) (0.048) (0.054)
Heal h (HEA) −0.011 −0.001 0.014 0.018
(0.050) (0.054) (0.050) (0.052)
Pe manen job (PJO) −0.001 0.031
(0.036) (0.036)
Age (AGE) −0.010*** −0.005***
(0.001) (0.001)
Educa ion (EDU) 0.117*** 0.087***
(0.032) (0.032)
Ma i al (MAR) 0.03 0.033
(0.032) (0.030)
Gende (GEN) −0.174*** −0.176***
(0.032) (0.034)
Occupa ion (OCC) 0.013*** 0.012***
(0.003) (0.003)
Obse a ions 1300 1300 1168 978 978 899
Pseudo R
2
0.034 0.035 0.1206 0.055 0.062 0.1322
X
2
59.40 60.14 171 63.62 70.48 118.48
Ma ginal e ec s. Robus s anda d e o s in pa en heses.
Sou ce: own elabo a ion.
***
p <0.01.
**
p <0.05.
*
p <0.1.
S. Fe n´
andez e al.
Science o he To al En i onmen 922 (2024) 171210
9
economic and socio- amilia easons, which opposes o he case o
Bangladesh in which he majo i y o he mig an s in he sample we e
men. This is aligned wi h he esul s exposed by La o e al. (2018),
which e eal an inc eased mobili y and independence among emale
mig an s in Ghana. Occupa ion (OCC) is signi ican oo in he explana-
ion o bo h kind o mig a ions in Ghana, which seems o be connec ed o
he idea o mig a ion as a esponse o a pa ial disequilib ium in labou
ma ke s. This way o unde s anding mig a ion ma ches he esul s by
Molini e al. (2016), which shows ha an his o ical mig a ion pa e n
linked o he demand o labou in indus ies such mining o ag icul u e
in speci ic a eas o he coun y has no changed signi ican ly since he
colonial pe iod. Finally, ha ing a pe manen job (PJO) o being ma ied
(MAR) does no seem o in luence mig a ion decisions in Ghana.
3.3. Mig a ion d i e s wi h en i onmen al s ess as mode a ing e ec :
econome ic esul s
As men ioned abo e, a p incipal componen analysis is ca ied ou o
deepen he esul s and hen new models a e es ima ed. The ini ial esul s
o he PCA o Bangladesh and Ghana a e p esen ed in Table A3 in
Appendix A, while he ac o loadings and unique a iances a e shown in
Table A4.
Fo he case o Bangladesh, Fac o 1 i sel will be he en i onmen al
s ess indica o . Howe e , o Ghana he en i onmen al s ess indica o
will be calcula ed as he a e age o Fac o 1 and Fac o 2. Using his
en i onmen al s ess a iable, we u he mo e explo e i s ole as a
mode a ing elemen on he o me con ol a iables. The esul s a e
shown in Table 6.
The esul s o Bangladesh a e obus o p e ious models. Fo he
case o mig aeco (economic mig a ion), none o he clima ic explana o y
a iables has an in luence on he p obabili y o mig a ing o economic
easons. Howe e , ha ing a pe manen job (PJO) and a highe educa-
ional le el (EDU) a ou economic mig a ion. On he o he hand, age
and being ma ied (MAR) educe he p obabili y o mig a ing o eco-
nomic easons. The main di e ence wi h espec o p e ious models
conce ns he occupa ion a iable (OCC), which is now signi ican when
explaining economic mig a ion. In addi ion, i is obse ed ha being
subjec ed o en i onmen al s ess posi i ely mode a es he occupa ion
d i e . The e o e, he occupa ion i sel bu also being a ec ed by en i-
onmen al s ess in luence he p obabili y o mig a ing.
In he model ega ding socio- amily mig a ions, no signi ican e ec s
a e ound wi h espec o he p e ious model wi hou in e ac ions.
Table 6
P obi models including en i onmen al s ess o Bangladesh and Ghana (model III).
Bangladesh Ghana
mig aeco mig asoci ami mig aeco mig asoci ami
(III) (III) (III) (III)
Housing (HOU) 0.035 0.025 −0.031 0.084
(0.047) (0.042) (0.052) (0.052)
Ecosecu i y (ECO) 0.037 0.082* 0.115* 0.071
(0.048) (0.043) (0.070) (0.069)
C op (CRO) 0.038 0.026 0.048 0.143*
(0.065) (0.056) (0.068) (0.074)
Wa e (WAT) −0.026 0.001 −0.032 0.053
(0.056) (0.051) (0.072) (0.071)
Foodsecu i y (FSE) −0.047 −0.024 −0.031 0.147*
(0.058) (0.052) (0.071) (0.077)
Heal h (HEA) 0.014 0.064 −0.125 −0.022
(0.063) (0.058) (0.082) (0.078)
Pe manen job (PJO) 0.162*** 0.139*** −0.122 −0.104
(0.033) (0.030) (0.082) (0.097)
Age (AGE) −0.015*** −0.012*** −0.010*** −0.002
(0.001) (0.001) (0.003) (0.003)
Educa ion (EDU) 0.101*** 0.015 0.007 0.027
(0.032) (0.029) (0.072) (0.071)
Ma i al (MAR) −0.273*** −0.254*** 0.143** 0.183***
(0.047) (0.050) (0.070) (0.062)
Gende (GEN) −0.180** −0.198***
(0.072) (0.075)
Occupa ion (OCC) 0.005* 0.003 −0.004 2.62E-04
(0.003) (0.003) (0.007) (0.007)
en i ons ess*Pe manen job (E-PJO) −0.047 −0.04 0.363 0.37
(0.032) (0.028) (0.231) (0.236)
en i ons ess*Age (E-AGE) 0.001 4.27E-04 0.003 −0.01
(0.001) (0.001) (0.008) (0.008)
en i ons ess*Educa ion (E-EDU) −0.007 0.003 0.361* 0.174
(0.032) (0.027) (0.212) (0.201)
en i ons ess*Ma i al (E-MAR) −0.038 −0.032 −0.377* −0.507**
(0.042) (0.035) (0.216) (0.202)
en i ons ess*Gende (E-GEN) 0.021 0.08
(0.220) (0.202)
en i ons ess*Occupa ion (E-OCC) 0.007** 0.005* 0.056*** 0.039**
(0.003) (0.003) (0.021) (0.020)
Obse a ions 1086 949 1168 899
Pseudo R
2
0.222 0.2187 0.1324 0.151
X
2
258.04 186.12 190.87 152.41
Ma ginal e ec s. Robus s anda d e o s in pa en heses.
Sou ce: own elabo a ion.
***
p <0.01.
**
p <0.05.
*
p <0.1.
S. Fe n´
andez e al.
